gsaltintas commited on
Commit
1fe3a86
·
verified ·
1 Parent(s): 55aeb76

Upload folder using huggingface_hub

Browse files
Files changed (5) hide show
  1. README.md +13 -11
  2. merges.txt +0 -0
  3. tokenizer.json +0 -0
  4. tokenizer_config.json +0 -0
  5. vocab.json +0 -0
README.md CHANGED
@@ -1,8 +1,10 @@
1
  ---
2
  license: mit
3
  language:
4
- - cmn #
5
- - jpn #['cmn_Hani', 'jpn_Jpan'] # ISO 639-3 code or "und" if not identifiable
 
 
6
  tags:
7
  - tokenizer
8
  - bpe
@@ -10,31 +12,31 @@ tags:
10
  - fineweb2
11
  ---
12
 
13
- # Byte-Level BPE Tokenizer: ['cmn_Hani', 'jpn_Jpan'] (16K)
14
 
15
- A **Byte-Level BPE** tokenizer trained on **['cmn_Hani', 'jpn_Jpan']** data from Fineweb-2-HQ.
16
 
17
  ## Training Details
18
 
19
  | Parameter | Value |
20
  |-----------|-------|
21
  | Algorithm | Byte-Level BPE |
22
- | Language | `['cmn_Hani', 'jpn_Jpan']` |
23
- | Target Vocab Size | 16,000 |
24
- | Final Vocab Size | 17,900 |
25
- | Pre-tokenizer | custom:jpn_Jpan |
26
  | Number handling | ltr_3digit |
27
  | Contraction handling | True |
28
  | Normalizer | NFC |
29
  | Special Tokens | `<s>`, `</s>`, `<pad>`, `<unk>` |
30
- | Training Shards | 4 |
31
 
32
  ## Usage
33
 
34
  ```python
35
  from transformers import AutoTokenizer
36
 
37
- tokenizer = AutoTokenizer.from_pretrained("flexitok/bpe_script_CmJp_16000")
38
  tokens = tokenizer.encode("Hello, world!")
39
  ```
40
 
@@ -47,4 +49,4 @@ tokens = tokenizer.encode("Hello, world!")
47
  ## Sample Encoding
48
  | Text | Tokens | Token IDs |
49
  |------|--------|-----------|
50
- | `Hello, world! 12345 This is a test. こんにちは` | `H, ell, o, ,, Ġw, orld, !, Ġ, 123, 45, ĠTh, is, Ġis, Ġa, Ġt, est, ., Ġ, ãģ, ĵ` | `42, 3452, 81, 14, 2927, 10693, 3, 223, 12071, 3557, 9838, 1476, 6231, 3791, 4071, 3540, 16, 223, 3577, 244` |
 
1
  ---
2
  license: mit
3
  language:
4
+ - dan
5
+ - deu
6
+ - nld
7
+ - swe #['dan_Latn', 'deu_Latn', 'nld_Latn', 'swe_Latn'] # ISO 639-3 code or "und" if not identifiable
8
  tags:
9
  - tokenizer
10
  - bpe
 
12
  - fineweb2
13
  ---
14
 
15
+ # Byte-Level BPE Tokenizer: ['dan_Latn', 'deu_Latn', 'nld_Latn', 'swe_Latn'] (32K)
16
 
17
+ A **Byte-Level BPE** tokenizer trained on **['dan_Latn', 'deu_Latn', 'nld_Latn', 'swe_Latn']** data from Fineweb-2-HQ.
18
 
19
  ## Training Details
20
 
21
  | Parameter | Value |
22
  |-----------|-------|
23
  | Algorithm | Byte-Level BPE |
24
+ | Language | `['dan_Latn', 'deu_Latn', 'nld_Latn', 'swe_Latn']` |
25
+ | Target Vocab Size | 32,000 |
26
+ | Final Vocab Size | 32,865 |
27
+ | Pre-tokenizer | custom:dan_Latn |
28
  | Number handling | ltr_3digit |
29
  | Contraction handling | True |
30
  | Normalizer | NFC |
31
  | Special Tokens | `<s>`, `</s>`, `<pad>`, `<unk>` |
32
+ | Training Shards | 8 |
33
 
34
  ## Usage
35
 
36
  ```python
37
  from transformers import AutoTokenizer
38
 
39
+ tokenizer = AutoTokenizer.from_pretrained("flexitok/bpe_script_Germ_32000")
40
  tokens = tokenizer.encode("Hello, world!")
41
  ```
42
 
 
49
  ## Sample Encoding
50
  | Text | Tokens | Token IDs |
51
  |------|--------|-----------|
52
+ | `Hello, world! 12345 This is a test. こんにちは` | `H, ello, ,, Ġworld, !, Ġ, 123, 45, ĠThis, Ġis, Ġa, Ġtest, ., Ġ, ãģ, ĵ, ãĤ, ĵ, ãģ«, ãģ` | `42, 13486, 14, 21745, 3, 223, 19219, 3832, 16775, 516, 270, 5190, 16, 223, 3768, 244, 5986, 244, 30698, 3768` |
merges.txt CHANGED
The diff for this file is too large to render. See raw diff
 
tokenizer.json CHANGED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json CHANGED
The diff for this file is too large to render. See raw diff
 
vocab.json CHANGED
The diff for this file is too large to render. See raw diff